You may have already followed our docker quickstart guide but in fact, the metrictank repository contains a variety of docker stacks for different purposes.
They can be started by cd'ing into their directory and running docker-compose up --force-recreate -V
Here's an overview. Note:
- not everything is detailed here. Plus this needs some more work. For specifics on any setup, please refer to the sources.
- every environment comes with grafana and cassandra unless otherwise noted
- every environment automatically provisions useful datasources and dashboards unless otherwise noted
plain, basic environment. uses metrictank docker image (with baked in binaries and configs), statsdaemon. Comes with 1 dashboard: the metrictank dashboard
Similar to docker-standard, but custom metrictank build is loaded, custom configs from scripts/config also: tag support, main+extra dashboards
Same as docker-dev but using scylladb instead of cassandra
Similar to docker-dev in terms of build and configuration. Also:
- kafka input and clustering backend
- short chunkspan & numchunks (benefits: easy to trigger mem and mem_and_cass requests, frequent cass saves, notifier messages etc)
- prometheus monitoring and jaeger tracing
- caddy-proxied endpooints/datasources to check with different org-id auth headers
- tsdb-gw
Similar to docker-dev-custom-cfg-kafka
- uses a cluster of 4 MT's (with 2x replication)
- separate graphite monitoring server
see our docker-cluster quickstart guide
Similar to docker-cluster but used for chaos testing
Used for testing against cosmosDB. Note: expects you to have a reachable cosmosdb and adjust settings accordingly. Does not auto-load datasources or dashboards.